North, Frederick

Objective Ambient listening with subsequent artificial intelligence enabled medical record documentation is changing how physicians interact with the electronic health record (EHR). We studied primary care physician use, documentation time, and changes in note metrics associated with ambient listening technology. Methods We calculated the percentage of physicians adopting the ambient listening tool, Abridge. Note input contributed by Abridge was determined by the percentage of total characters in generated notes. Using the Epic EHR Signal metrics, we analyzed the change in note documentation time and note length for Abridge users before and after they implemented Abridge in their note generation process. Results The percentage of physicians using Abridge increased from 15% to 50% in 8 weeks. Adoption was also reflected in the percentage of the note attributable to the use of ambient listening technology. Note content generated by Abridge tripled, from 5% to 15% over the same 8 weeks. After Abridge adoption, 332 primary care physicians had a mean time of 4.16 minutes [confidence interval (CI) 95%; 3.86, 4.46] time in notes per note in a paired comparison to 5.11 minutes [CI 95%; 4.75, 5.47] per note prior to Abridge. Mean time saved per note after Abridge was 0.95 minutes [CI 95%; 0.48, 1.42, p < 0.0001], an 18.6% reduction in time. Mean note length for Abridge adopters increased by 238 characters [CI 95%; 89, 387, p < 0.0001] from a pre-implementation mean of 4404 characters per note (+5.4%). Conclusion Ambient listening had rapid uptake and widespread use by physicians in primary care. It was also associated with a significant decrease in time in notes.

Background: Patients performing self-triage for chest pain need to pick an algorithm which includes questions pertinent to chest pain to triage accurately. Objective: Our study reviews patient self-triage choices and outcomes for patients who completed an online self-triage encounter before being triaged by a nurse for chest pain. Methods: Patients who underwent telephone nurse triage for chest pain, and also had an online self-triage encounter within the prior 24 hours were reviewed for the frequency with which they chose “chest pain” as their symptom to triage for during self-triage, whether they reported chest pain during the self-triage encounter (if asked as part of the algorithm for their chosen self-triage symptom) and whether the patient had any follow up evaluation within 1 week. Results: There were 70 self-triage and nurse triage dyads during the study period where triage nurses chose “chest pain” as the primary symptom to be triaged for during the telephone triage encounter. Of these, only 5 (7.1%) patients chose the “chest pain” self-triage option during online self-triage. During the self-triage encounter, 50 (71%) reported chest pain, 10 (14%) denied chest pain and 10 chose to self-triage their symptoms with an algorithm that did not include a question on whether chest pain was present. Of the 70 total dyads 59 (84%) had a follow up emergency department (40 patients) or office visit (19 patients) documented. Conclusion: Most patients chosen to be triaged for chest pain by triage nurses did not choose the online self-triage algorithm for chest pain even though the majority did report chest pain during their self-triage encounter (when asked). This is concerning from a safety perspective as a non-chest pain self-triage option chosen by the patient may not always ask about the presence or absence of chest pain and thus could increase the risk of being triaged inaccurately. Implications for improvement of self-triage could include diagrams for patients to choose the symptom area.

Introduction: Online self-triage/symptom checkers are a relatively new e-Health modality. We aimed to evaluate the degree to which patient’s chosen symptom during the use of an online self-triage tool was congruent with that chosen by a nurse performing phone triage. Methods: Patient encounters from September 1, 2022, to September 30, 2023, were reviewed for patients who completed both online self-triage and a subsequent nursing phone triage within 24 hours. The symptoms chosen by the patient for their self-triage encounter and the symptoms chosen by the nurse during the subsequent nurse triage were reviewed. The symptoms chosen in these two triage encounters were then compared for congruency. A subset of our database was evaluated for congruency between the self-triage and the nurse phone triage endpoint recommendations. Results: There were 5,443 encounters meeting inclusion criteria to study for congruency between self-triage and nurse triage symptom choice. The range of time between self-triage and nurse triage ranged from 0 min to 1,439 min (23.9 hours) with a median of 23 minutes between self and nurse triage encounters. Symptom congruency between the two methods found 74.2% (4,038) to be completely congruent while 989 (18.2%) were somewhat congruent. Only 7.6% (416) had no apparent congruency. Out of 1,705 self and nurse triage dyads eligible for study for agreement between endpoint recommendations, 244 (14.3%) were under-triaged, 1,110 (65.1%) were triaged to the same level of care and 351 (20.6%) were over-triaged. Conclusion: Congruency between symptoms chosen by the patient for triage and nurse triage were highly congruent with only 7.6% of the sample showing no apparent congruency between the self-triage symptom and the nurse triage symptom. When compared to nurse triage, most self-triage endpoint recommendations were either triaged to the same level of care or were over-triaged to a higher level of care.

Introduction: Online self-scheduling of medical appointments is increasingly common. An automated waitlist can be used for patients who desire an earlier appointment time if one becomes available after they are scheduled. Our study examines outcomes of an automated waitlist and self-rescheduling process. Methods: We studied outcomes of an automated waitlist self-rescheduling process in which patients with existing appointments elected to be placed on an automated waitlist for an earlier appointment offer. When software found earlier dates for the same visit type, patients were then notified through an automated process and could self-reschedule. We reviewed appointments for which patients were sent new offers when earlier appointment slots were found. We compared the accepted appointment offers with the original scheduled appointments and determined the number of days that the appointment had been moved up. Results: Spanning the calendar year 2023 there were 1 019 698 appointment offers generated by an automated waitlist process for 229 998 appointments and sent to 164 248 patients. The waitlist process automatically found open appointments as they became available and sent the first new appointment offer within 2 days after being placed on the waitlist for 74 736 (32.5%) of the 229 998 waitlisted appointments. Patients sent back at least 1 response for 104 554 (45.4%) of the waitlisted appointments. Of the responses, 56 636 accepted one of the sent offers for an accept rate of 24.6% (56 636/229 998). For accepted, moved-up visits, appointments were self-rescheduled earlier by a mean of 22.6 days (95%CI; 22.2, 22.9, P < .0001). Conclusion: New appointments can be successfully self-rescheduled using an automated waitlist process that allows patients to accept or decline new appointment offers. This process can increase the efficiency of scheduling and decrease appointment wait time for patients desiring more timely access to healthcare. In addition, this process can be successfully applied across several different appointment type categories.

Introduction Patient self-scheduling of medical appointments is becoming more common in many medical institutions. However, the complexity of scheduling multiple specialties, following scheduling guidelines, and managing appointment access requires a variety of processes for a diverse inventory of self-schedulable appointment types. Methods From 7 unique patient self-scheduling methods, we captured counts of successfully self-scheduled and completed appointments. A process map was created to show the paths of 5 different primary self-scheduling processes (new appointment self-scheduling) and 2 secondary self-scheduling processes (existing appointment self-rescheduling). Results There were 7 unique processes that led to 733,651 successfully self-scheduled completed visits from January 1 to December 31, 2023 at a multisite, multispecialty clinic. The self-scheduling processes consisted of the following: (1) Ticket offer (appointment “ticket” offers for specific visits generated by a provider order or system rules), the software “ticket” sent to the patient permits “admission” to self-schedule calendar templates (341,591 uses, 46.6%); (2) direct self-scheduled visit for prequalified visit types (203,593 uses, 27.6%); (3) self-reschedule option (patient option to reschedule existing appointment, 79,706 uses, 10.9%); (4) new patient self-scheduled visit via clinic website (does not require portal access, 54,367 uses, 7.4%). (5) automated waitlist self-rescheduled visit (38,649 uses, 5.3%); (6) automated waitlist self-scheduled visit of previously unscheduled visit (10,939 uses, 1.5%); and (7) self-triage self-scheduled visit (4806 uses, 0.7%). Conclusion The processes for self-scheduling are expanding. Our multispecialty clinic has implemented 7 different processes to help patients successfully self-schedule medical appointments. Some of the processes occur before initial scheduling (such as self-triage), and some are implemented after successful scheduling has already occurred (self-rescheduling option and self-rescheduling aided by an automated waitlist). Continued research is needed to look for measures of success beyond the ability to complete a self-scheduled visit, including the accuracy of the booking (right provider, location, and length of visit).

Background Self-scheduling of medical visits is becoming more common but the complexity of applying multiple requirements for self-scheduling has hampered implementation. Mayo Clinic implemented self-scheduling in 2019 and has been increasing its portfolio of self-schedulable visits since then. Our aim was to show measures quantifying the complexity associated with medical visit scheduling and to describe how opportunities and challenges of scheduling complexity apply in self-scheduling. Methods We examined scheduled visits from January 1, 2022, through August 24, 2023. For seven visit categories, we counted all unique visit types that were scheduled, for both staff-scheduled and self-scheduled. We examined counts of self-scheduled visit types to identify those with highest uptake during the study period. Results There were 9555 unique visit types associated with 20.8 M (million) completed visits. Self-scheduled visit types accounted for 4.0% (838,592/20,769,699) of the completed total visits. Of seven visit categories, self-scheduled established patient visits, testing visits, and procedure visits accounted for 93.5% (784,375/838,592) of all self-scheduled visits. Established patient visits in primary care (10 visit types) accounted for 273,007 (32.6%) of all self-scheduled visits. Testing visits (blood and urine testing, 2 visit types) accounted for 183,870 (21.9%) of all self-scheduled visits. Procedure visits for screening mammograms, bone mineral density, and immunizations (8 visit types) accounted for 147,358 (17.6%) of all self-scheduled visits. Conclusion Large numbers of unique visit types comprise a major challenge for self-scheduling. Some visit types are more suitable for self-scheduling. Guideline-based procedure visits such as screening mammograms, bone mineral density exams, and immunizations are examples of visits that have high volumes and can be standardized for self-scheduling. Established patient visits and laboratory testing visits also can be standardized for self-scheduling. Despite the successes, there remain thousands of specific visit types that may need some staff-scheduler intervention to properly schedule.

Background Self-scheduling of medical visits is becoming available at many medical institutions. We aimed to examine the self-scheduled visit counts and rate of growth of self-scheduled visits in a multispecialty practice. Methods For 85 weeks extending from January 1, 2022 through August 24, 2023, we examined self-scheduled visit counts for over 1500 self-scheduled visit types. We compared completed self-scheduled visit counts to all scheduled completed visit counts for the same visit types. We collected counts of the most frequently self-scheduled visit types for each week and examined the change over time. We also determined the proportion that each visit type was self-scheduled. Results There were 20,769 699 completed visits during the course of the study that met the criteria for inclusion. Self-scheduled visits accounted for 4.0% of all completed visits (838 592/20,769 699). Over the 85-week span, self-scheduled visits rose from 3.0% to 5.3% of the total. There were 1887 unique visit types that were associated with completed visits. There were just 6 appointment visit types of the total 1887 self-scheduled visit types that accounted for 50.7% of the total 838 592 self-scheduled visits. Those 6 visit types were a lab blood test visit (19.5%, 163 K visits), two Family Medicine office visit types (13.0%, 109 K visits), a screening mammogram visit type (6.6%, 55 K visits), a scheduled express care visit type (6%, 50 K visits) and a COVID immunization visit type (5.7%, 48 K visits). Twenty-one visit types that were self-scheduled accounted for 75% of the total self-scheduled visits. Four seasonal visits, accounting for 10.6% of the total self-scheduled visits, were responsible for almost all the non-linear change in self-scheduling. Conclusion Self-scheduling accounted for a small but growing percent of all outpatient scheduled visits in a multispecialty, multisite practice. A wide range of visit types can be successfully self-scheduled.

Introduction Online symptom checkers are a way to address patient concerns and potentially offload a burdened healthcare system. However, safety outcomes of self-triage are unknown, so we reviewed triage recommendations and outcomes of our institution's depression symptom checker. Methods We examined endpoint recommendations and follow-up encounters seven days afterward during 2 December 2021 to 13 December 2022. Patients with an emergency department visit or hospitalization within seven days of self-triaging had a manual review of the electronic health record to determine if the visit was related to depression, suicidal ideation, or suicide attempt. Charts were reviewed for deaths within seven days of self-triage. Results There were 287 unique encounters from 263 unique patients. In 86.1% (247/287), the endpoint was an instruction to call nurse triage; in 3.1% of encounters (9/287), instruction was to seek emergency care. Only 20.2% (58/287) followed the recommendations given. Of the 229 patients that did not follow the endpoint recommendations, 121 (52.8%) had some type of follow-up within seven days. Nearly 11% (31/287) were triaged to endpoints not requiring urgent contact and 9.1% (26/287) to an endpoint that would not need any healthcare team input. No patients died in the study period. Conclusions Most patients did not follow the recommendations for follow-up care although ultimately most patients did receive care within seven days. Self-triage appears to appropriately sort patients with depressed mood to emergency care. On-line self-triaging tools for depression have the potential to safely offload some work from clinic personnel.

Background Patients often present to emergency departments (EDs) with concerns that do not require emergency care. Self-triage and other interventions may help some patients decide whether they should be seen in the ED. Symptoms associated with low risk of hospitalization can be identified in national ED data and can inform the design of interventions to reduce avoidable ED visits. Methods We used the National Hospital Ambulatory Medical Care Survey (NHAMCS) data from the United States National Health Care Statistics (NHCS) division of the Centers for Disease Control and Prevention (CDC). The ED datasets from 2011 through 2020 were combined. Primary reasons for ED visit and the binary field for hospital admission from the ED were used to estimate the proportion of ED patients admitted to the hospital for each reason for visit and age category. Results There were 221,027 surveyed ED visits during the 10-year data collection with 736 different primary reasons for visit and 23,228 hospitalizations. There were 145 million estimated hospitalizations from 1.37 billion estimated ED visits (10.6%). Inclusion criteria for this study were reasons for visit which had at least 30 ED visits in the sample; there were 396 separate reasons for visit which met this criteria. Of these 396 reasons for visit, 97 had admission percentages less than 2% and another 52 had hospital admissions estimated between 2% and 4%. However, there was a significant increase in hospitalizations within many of the ED reasons for visit in older adults. Conclusion Reasons for visit from national ED data can be ranked by hospitalization risk. Low-risk symptoms may help healthcare institutions identify potentially avoidable ED visits. Healthcare systems can use this information to help manage potentially avoidable ED visits with interventions designed to apply to their patient population and healthcare access.

Background Although online self-triage is easily accessible, little is known about the patients who use self-triage or their subsequent diagnoses. We compared ear/hearing self-triage subsequent diagnoses to ear/hearing visit diagnoses in emergency departments (ED) and ambulatory clinics across the United States. Methods We compared International Classification of Diseases version 10 (ICD10) coded diagnoses following online self-triage for ear/hearing concerns with those from national ED and ambulatory clinic samples. We used data from the Centers for Disease Control (CDC) National Hospital Ambulatory Medical Care Survey (NHAMCS) and National Ambulatory Medical Care Survey (NAMCS) for comparison. Using matched ear/hearing diagnostic categories for those aged 1 and over, we compared self-triage diagnosis frequencies with national ED and ambulatory diagnosis frequencies. Results Following ear/hearing self-triage, there were 1092 subsequent office visits with a primary diagnosis code. For five frequently diagnosed ear/hearing conditions (i.e., suppurative and nonsuppurative otitis media [OM], otalgia, otitis externa, and cerumen impaction), there was a strong correlation between diagnosis counts made following self-triage and estimated counts of national ED visit diagnoses (r = 0.94; CI 95% [0.37 to 0.99]; p = .016, adjusted r2 = 0.85). Seven diagnoses were available to compare with the national ambulatory sample; correlation was r = 0.79; CI 95% [0.08 to 0.97]; p = .037, adjusted r2 = 0.54. For ages 1 and over, estimated hospital admissions from the national ED visits for ear/hearing were 0.76%, CI 95% [0.28-2.1%]; estimated total national ear/hearing ED visits were 7.5 million (for 4 years, 2016 through 2019). Conclusion The strong correlation of ear-related self-triage diagnoses with national ED diagnoses and the low hospitalization risk for these diagnoses suggests that there is an opportunity for self-triage of ear/hearing concerns to decrease ED visits for these symptoms.